Computer-assisted assessment of human sperm morphology: usefulness in predicting fertilizing capacity of human spermatozoa

The usefulness of sperm morphology to predict the outcome of human sperm fertilizing capacity was examined. Semen samples from 50 male patients attending the infertility clinic of a tertiary referral institution were studied. Sperm morphology was classified both by visual assessment and computer-assisted image analysis. In addition, morphometric analysis of the spermatozoa was measured by the morphologizer. Multivariate discriminant analysis was used to evaluate the usefulness of these morphology parameters for predicting the outcome of the zona-free hamster oocyte sperm penetration assay. The manually derived percent of spermatozoa with normal and small head were selected to be of discriminating value in predicting the outcome of the zona-free hamster oocyte penetration test. The accuracy of correctly classifying the outcome of zona-free hamster oocyte penetration test by these two parameters in combination was 84%, whereas assessment of sperm morphology with morphometric analysis by the morphologizer selected a total of eight variables, which together predicted sperm fertilizing capacity with 74% accuracy. Addition of the morphologizer-derived parameters to those derived manually did not significantly improve the predictive value. We conclude that the results of the zona-free hamster egg penetration test could be predicted using manual assessment of sperm morphology and computer-assisted morphometric analysis did not add further information.